Electronic coupling of multiexcitonic states: Development of quantum chemical methods and their application to singlet fission and triplet fusion

传统的光伏装置仅收集入射太阳光的一部分。由于可再生能源的重要性，已经投入大量的努力来提高太阳能电池的效率。两种密切相关的策略，用于收集那些部分的太阳光谱，这通常是失去了热转换，是两个低能量的三重态激子的融合和上转换为较高能量的单重态激子（三重态-三重态湮灭上转换，TTA-UC）和单重态裂变（SF），分裂的高能量的单重态激子到两个较低能量的三重态激子。尽管关于TTA-UC和SF的一般机制存在共识，但细节仍然存在大量争议，并且是一个活跃的研究领域。例如，事实证明，更快的SF过程不一定更有效。此外，偏离自旋统计极限的量子产率已被观察到在几个SF和TTA-UC系统。这意味着（T：T）三重态的形成不遵循自旋统计，或者精细结构相互作用耦合了（T：T）态。这里提出的项目设置在理论和计算化学领域。三重态对是多激子态，即，它们相对于电子基态被双重激发，这阻止了用于描述激发能量转移的Förster和Dexter型模型的应用。拟议的研究的主要目标是把强有力的工具的自旋允许和自旋禁止耦合的多激子状态的计算评估超出最小的活性空间模型，并将这些工具应用到一系列具有挑战性的案例研究共价连接的聚并苯二聚体或紧密结合的遭遇复合物的聚并苯。在这里，我们的目标是了解，从而得到一个处理的因素和结构参数，引导在这样的系统中的SF和TTA-UC的效率。特别是，我们打算设计和参数化一个新的半经验DFT/MRCI哈密顿容忍一个粒子的碎片本地化，非正则轨道的基础上，从而通过DFT/MRCI本征态的diabatization的非绝热耦合矩阵元素的确定容易。此外，我们将开发计算机代码和近似，使计算的单重态，三重态和五重态之间的电子自旋-自旋耦合在大分子系统。这些代码的应用将提供访问上述（T：T）状态之间的精细结构相互作用的大小。这些现象的复杂性，分子系统的庞大规模和潜在的电子状态的不同性质-特别是多激子状态-代表任何给定的激发态电子结构方法的真实的挑战，但我们有信心，我们可以解决这些问题，在本项目。

Conventional photovoltaic devices harvest only fractions of the incident solar light. Owing to the importance of renewable energy, substantial effort has been put into improving the efficiency of solar cells. Two closely related strategies for harvesting those parts of the solar light spectrum, which are typically lost by heat-conversion, are the fusion of two lowenergy triplet excitons and upconversion to a higher-energy singlet exciton (triplet–triplet annihilation up conversion, TTA-UC) and singlet fission (SF), the fission of a high-energy singlet exciton into two lower-energy triplet excitons. Although there is consensus regarding the general mechanisms of TTA-UC and SF, the details are still heavily debated and an active field of research. It turned out, for example, that the faster SF process is not necessarily the more efficient one. Moreover, deviations of the quantum yield from the spin-statistical limit have been observed in several SF and TTA-UC systems. This means either that the formation of the (T: : :T) triplet pair states do not follow spin statistics or that fine-structure interactions couple the (T: : :T) states. The here proposed project is set in the field of theoretical and computational chemistry. The triplet-pair states are multiexcitonic states, i.e., they are doubly excited with respect to the electronic ground state, which prevents the application of Förster- and Dexter-type models for describing the excitation energy transfer. The main objectives of the proposed research are to put in place powerful tools for the computational evaluation of spin-allowed and spin-forbidden couplings of multiexcitonic states beyond minimal active space models and to apply these tools to a series of challenging case studies on covalently linked polyacence dimers or tightly bound encounter complexes of polyacences. Herein, we aim to understand and thereby get a handle on the factors and structural parameters that steer the efficiency of SF and TTA-UC in such systems. In particular, we intend to design and parameterize a new semiempirical DFT/MRCI Hamiltonian which tolerates a one-particle basis of fragment-localized, noncanonical orbitals, thus easing the determination of nonadiabatic coupling matrix elements via diabatization of the DFT/MRCI eigenstates. Moreover, we will develop computer codes and approximations enabling the calculation of electronic spin–spin coupling between singlet, triplet, and quintet states in large molecular systems. Application of these codes will provide access to the magnitude of fine-structure interactions between the aforementioned (T: : :T) states. The complexity of these phenomena, the sheer size of the molecular systems and the different natures of the electronic states potentially involved — in particular of multiexcitonic states — represent real challenges for any given excited-state electronic structure method but we are confident that we can tackle these problems in the present project.